Proposed as processing of matching a large volume of data from numerous vehicles with a map at a high speed are the method of efficiently reducing a volume of event data (see Patent Literature 1) and the method of efficiently reducing road candidates to be matched (see Patent Literature 2). Event data is data indicative of a position and a state of a vehicle.
Recited in Patent Literature 1 is the method of narrowing down, on the side of a vehicle as an event generating source, data to be uploaded to a server. In the narrowing-down method recited in Patent Literature 1, with a configuration such as a radius of curvature of a road held as data, a sampling rate is controlled. In a case of traveling on a straight road, for example, the volume of data is suppressed by lowering a sampling rate.
In Patent Literature 2, recited as a latter case is the method, with roads classified into a highest layer (main road network), a middle layer (network of road whose width is not less than 5.5 m) and a lowest layer (all the road networks), of executing matching starting with the highest layer. In the method recited in Patent Literature 2, when a matching result fails to satisfy a certain condition, speed-up is realized by trying matching with a further lower layer.
Recited in Patent Literature 3 is the navigation system using mesh data obtained by dividing a predetermined geographical range into a plurality parts. Recited in Patent Literature 4 is the system for calculating the degree of traffic congestion by calculating an average vehicle speed.
Patent Literature 1: Japanese Patent Laying-Open No. 2004-280521.
Patent Literature 2: Japanese Patent Laying-Open No. 2004-354395.
Patent Literature 3: Japanese Patent Laying-Open No. 2004-177364.
Patent Literature 4: Japanese Patent Laying-Open No. 2005-301643.
The method recited in Patent Literature 1 can be considered to be a method which can be realized by making use of event data from a vehicle aiming at more precisely generating a traveling track of each vehicle. Such a method, however, disables use of event data for other purpose. In a case of estimating traffic congestion information from event data, for example, while each vehicle is allowed to control a sampling rate after recognizing a relationship with surrounding vehicles, the method recited in Patent Literature 1 is allowed to calculate a sampling rate only from fixed information such as a radius of curvature of a road. The method recited in Patent Literature 1 is therefore not applicable to estimation of traffic congestion information.
It is preferable to be usable not only for the specific object of precisely comprehending a traveling track but also for various kinds of analyses related to traffic conditions. Moreover, even when event data is transmitted from a large number of vehicles (e.g. several tens of thousands), it is preferable to correlate event data and roads at a high speed.
With the method recited in Patent Literature 2, the lower the layer becomes, the less the effect of speeding-up becomes, so that when matching with a lowest layer, the effect of speeding-up cannot be obtained. Accordingly, when making use of event data of a vehicle in a lower layer, no effect of speeding-up is obtained. When analysis is made for estimating an arrival time on a long-distance path, for example, because a vehicle uses a main road, high-speed processing is enabled by the method recited in Patent Literature 2, while when estimating an arrival time of a vehicle running on a daily road or the like, the effect of speed-up cannot be obtained.
Under these circumstances, the present invention aims at providing a general-purpose map matching system, a map matching method, a map matching device applicable to such a map matching system, a road data generating device, a map matching program and a road data generating program which enable high-speed map matching while maintaining a precision of an analysis based on a map matching result even when event data is transmitted from a large number of vehicles and which are not limited to a specific purpose.